Experimental Gerontology 126 (2019) 110706

Contents lists available at ScienceDirect

Experimental Gerontology
journal homepage: www.elsevier.com/locate/expgero

Hemoglobin levels and low bone mineral density in non-anemic older T

adults: Secondary analysis of the Korean National Health and Nutrition
Examination Survey☆
⁎
Hyoung-Sik Kima, Hye-Min Parkb, Hye Sun Leec, Yong-Jae Leeb,
a
Department of Orthopedic Surgery, Yonsei University College of Medicine, Seoul, Republic of Korea
b
Department of Family Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea
c
Biostatistics Collaboration Unit, Yonsei University College of Medicine, Seoul, Republic of Korea

A R T I C LE I N FO A B S T R A C T

Section Editor: Emanuele Marzetti Background: Although, previous studies have reported a positive association between hemoglobin levels and
Keywords: bone mineral density (BMD), the majority of the studies were limited in patients with chronic hypoxemic
Hemoglobin conditions and findings concerning the association among non-anemic populations are inconclusive. We aimed
Bone mineral density to examine the association between hemoglobin levels and BMD in non-anemic healthy adults.
Osteopenia Methods: This cross-sectional study included 3626 non-anemic men and women aged ≥ 60 years who partici-
Osteoporosis pated in the Korean National Health and Nutrition Examination Survey (KNHANES). The BMD of the lumbar
spine and both femurs was measured by dual-energy X-ray absorptiometry (DXA). Participants with T-score for
BMD < −1.0 SD were defined as having low BMD. The odds ratios (ORs) and 95% confidence intervals (CIs) for
low BMD were calculated using multiple logistic regression analyses across sex-specific hemoglobin quartiles.
Results: The prevalence of low BMD gradually decreased in accordance with hemoglobin quartiles in both sexes.
Compared with the group in the lowest quartile, the OR (95% CI) for low BMD in the lumbar spine was 0.78
(0.54–0.93) for men and 0.67 (0.50–0.93) for women after adjusting for age, BMI, tobacco smoking, alcohol
intake, physical activity, walking difficulty, household income, total calorie intake, calcium intake, iron intake,
25(OH)D, alkaline phosphatase, and parathyroid hormone levels. However, these positive associations were not
found in femur after adjusting for the same co-variables.
Conclusions: Hemoglobin levels were inversely associated with low BMD in lumbar spine among non-anemic
adults.

1. Introduction nutrition, physical activity, education level, prolonged use of gluco-

Osteoporosis is a metabolic bone disease characterized by micro-
architectural deterioration, decreased bone mass strength with a con-
sequent higher risk of fragility fracture, which lead to an increase in
disability, hospital cost, and mortality (Papadimitriou et al., 2017).
According to the World Health Organization (WHO), osteoporosis is
estimated to affect over 200 million people worldwide and the pre-
valence of osteoporosis in Korean adults increased during the last
decade from 6.1% to 13.1% for men and from 24.3 to 35.5% for women
(Ha, 2016).
Bone mineral density (BMD) is affected by various factors related to
calcium homeostasis including age, sex, family history, menopause,
corticoids, smoking and alcohol intake. (Reginster and Burlet, 2006).
Given the consequent higher risk of fragility fractures and the global
burden of disease, early identification of potentially modifiable factors
related to low BMD is important from a public health perspective.
Several biomarkers have been suggested to be useful predictors of
low bone mass including serum calcium, 25-hydroxyvitamin D (25(OH)
D), and ferritin. Moreover, previous studies have suggested that anemia
or relatively low hemoglobin levels, even those within the normal
range, are associated with BMD and an increase in fracture risk; how-
ever, these associations remain controversial (Salive et al., 1992; Leng
et al., 2002; Duh et al., 2008). Indeed, the majority of previous studies
that have looked into the relationship between hemoglobin levels and

☆
Sources of support for research: None.
⁎
Corresponding author at: Department of Family Medicine, Yonsei University College of Medicine Gangnam Severance Hospital, 211 Eonju-ro, Gangnam-gu, Seoul,
Republic of Korea.
E-mail address: ukyjhome@yuhs.ac (Y.-J. Lee).

https://doi.org/10.1016/j.exger.2019.110706
Received 17 June 2019; Received in revised form 1 August 2019; Accepted 19 August 2019
Available online 20 August 2019
0531-5565/ © 2019 Elsevier Inc. All rights reserved.
H.-S. Kim, et al.
BMD have been based on non-axial peripheral BMD measurement
methods, whereas the International Society for Clinical Densitometry
(ISCD) and the WHO recommend central DXA in axial bone such as
lumbar spine, total hip, or femoral neck (Organization, 1968). Likewise,
the study populations used in most of the previous studies on this area
of research have been limited to Western populations with chronic
hypoxemic medical conditions such as chronic obstructive pulmonary
disease (Rutten et al., 2013), renal failure (Taal et al., 1999), and an-
emia (Sarrai et al., 2007).
In the present study, we investigated the association between he-
moglobin levels within the normal range and central bone mineral
density as measured by DXA in apparently healthy Korean men and
women based on the 2009–2010 Korean National Health and Nutrition
Examination Survey (KNHANES).
2. Materials and methods
2.1. Study population
This study was based on data obtained from the 2009 and 2010
KNHANES, a nationally representative survey conducted by the Korean
Ministry of Health and Welfare. The target population of the survey
included non-institutionalized civilians in Korea who were over one
year of age. Sampling units consisted of households selected through a
stratified, multistage, probability-sampling design based on geographic
area, sex, and age group using household registries. Sampling weights,
which were representative of the probability of being sampled, were
assigned to each participant to ensure that the results represented the
overall Korean population. Participants completed a four-part survey
that consisted of a health interview, health behavior questionnaire, a
health examination, and a nutrition questionnaire.
In this study, 5608 participants aged ≥ 60 years old were included
in the 2009 and 2010 KNHANES. We excluded participants who met at
least one of the following criteria: those with a history of anemia, is-
chemic heart disease, stroke, thyroid, hepatobiliary, chronic renal dis-
ease, or cancers; those with hemoglobin level < 13 g/dL in men, <
12 g/dL in women; those with missing data for BMD (T-score), he-
moglobin, alkaline phosphatase, parathyroid hormone, and 25(OH)D
levels; those receiving hormone replacement therapy; those taking vi-
tamin and mineral supplements. Following these exclusions, a total of
3626 Koreans (1673 men, 1953 women) were in included in the final
analysis after exclusions. The KNHANES was approved by the
Institutional Review Board of the Korea Centers for Disease Control and
Prevention (Approval numbers: 2009-01CON-03-2C and 2010-02CON-
21-C). In addition, this study complied with the ethical principles of the
Declaration of Helsinki.
2.2. Data collection
For the 2009 and 2010 KNHANES, citizens were informed that they
had been randomly selected as a household to voluntarily participate in
a nationally representative survey conducted by the Korean Ministry of
Health and Welfare, and that they had the right to refuse participation
in accordance with the National Health Enhancement Act supported by
the National Statistics Law of Korea. All participants gave written in-
formed consent to participate in the study. The Korea Centers for
Disease Control and Prevention also obtained written informed consent
to use blood samples from participants for further analysis. The health
examinations, which were performed in 2009 and 2010, included a
medical history, physical examination, a questionnaire about health-
related behaviors, anthropometric measurements, and biochemical in-
dices.
Physical examinations were performed by trained medical staff
following standardized procedures. Body mass and height were mea-
sured to the nearest 0.1 kg and 0.1 cm, respectively, in light indoor
clothing, without shoes. Body mass index (BMI) was calculated as the
2
Experimental Gerontology 126 (2019) 110706
ratio of weight (kg) to height2 (m2). BMI was categorized into three
groups, namely, underweight (BMI < 18.5 kg/m2), normal (BMI
18.5–24.9 kg/m2), and obese (BMI ≥ 25.0 kg/m2) according to the
Korean Society for the Study of Obesity. Participants were asked about
lifestyle behaviors, including cigarette smoking and alcohol consump-
tion. Smoking status was categorized as a current smoker (a person who
smokes cigarettes daily) or not-current smoker (a person who has never
smoked or who smoked in the past but who does not currently smoke
cigarettes). Alcohol consumption was assessed into two categories
based on how often participants consumed any type of alcohol, namely,
current drinker (a person who drinks alcohol, even once a month) or
not-current drinker (a person who drinks less than once a month).
Participants completed the International Physical Activity
Questionnaire (IPAQ) to determine the frequency of physical activity.
Household income was classified into four quartiles from the lowest to
the highest. Dietary intake was assessed by administrating a Food
Frequency Questionnaire (FFQ) using a single 24-h recall method.
Before testing, all participants were instructed to maintain their usual
dietary habit. Total calorie intake, calcium intake, and iron intake were
calculated by Can-Pro 2.0, computerized nutrient intake assessment
software developed by the Korean Nutrition Society. If participants
were being treated for any disease, they were asked for data regarding
their diagnosis as well as a list of medications currently being taken.
Completed questionnaires were reviewed by trained staff and entered
into a database. Participants were also divided into two groups ac-
cording to the degree of physical activity, either difficulty in walking or
no difficulty in walking.
Serum hemoglobin levels were measured using an SLS hemoglobin
(no cyanide) method with an XE-2100D instrument (Sysmex, Tokyo,
Japan). Serum 25(OH)D concentrations were measured with a radio-
immunoassay (Diasorin Inc., Stillwater, MN, USA) using a γ-counter
(1470 Wizard; PerkinElmer, Turku, Finland). The coefficient of varia-
tion of control material was ≤7.6% for low level and ≤5.8% for high
level, respectively. The central testing institute has participated in the
international Vitamin D External Quality Assessment Scheme (DEQAS).
Results of the DEQAS proficiency testing showed less than ± 2.0 stan-
dard deviation index (SDI), except the SDI range from −2.40 to +2.87
due to random error in 2010, without negative or positive trend. Serum
parathyroid hormone (PTH) levels were measured using the N-tact PTH
assay with LIAISON (Diasorin Inc., Stillwater, MN, USA), which is a
chemiluminescence immunoassay method.
The BMD (g/cm2) measurements of central skeletal sites (femoral
neck, and lumbar spine) were obtained using DXA (QDR 4500A;
Hologic Inc., Waltham, MA, USA). BMD measurements were performed
according to the 2007 ISCD recommendations. The DXA instruments
were calibrated as previously described (Schoeller et al., 2005), and
reference values were obtained using this method (Kelly et al., 2009).
DXA calibration was performed by measuring phantoms, which are
reference standards, on a daily basis. The results were analyzed using
standard techniques and Korean Society of Osteoporosis and Hologic
Discovery software (version 13.1). L1–4 values were chosen for the
determination of lumbar spine BMD. When these specific vertebrae
were not suitable for analysis due to compression fracture, degenerative
changes or any other reason, the calculation of BMD excluded the af-
fected vertebrae. Participants with a BMD T-score of < −1.0 standard
deviation (SD) were placed in the low BMD group.
2.3. Statistical analysis
Because hemoglobin levels differ significantly by sex, hemoglobin
quartiles were categorized separately as follows: Q1, ≤14.2; Q2,
14.4–14.8; Q3, 15.0–15.6; and Q4, ≥15.7 g/dL for men and Q1, ≤12.7;
Q2, 12.8–13.2; Q3, 13.2–13.8; and Q4, ≥13.9 g/dL for women. In this
survey, the sample weights were constructed for sample participants to
represent the Korean population by accounting for the complex survey
design, survey non-response and post-stratification. Therefore, in
 H.-S. Kim, et al.

Table 1
The demographic and biochemical characteristics of the subjects according to hemoglobin quartiles.
Q1 ≤ 14.2 Q2 14.3–14.8 Q3 14.9–15.6 Q4 ≥ 15.7 P-valuea Post-hocb Q1 ≤ 12.7 Q2 12.8–13.2 Q3 13.3–13.8 Q4 ≥ 13.9 P-valuea Post-hocb

n 467 372 457 377 425 437 452 486

Age (year) 67.9 ± 0.5 65.8 ± 0.4 64.8 ± 0.3 64.4 ± 0.4 < 0.001 a,b,c 68.1 ± 0.5 67.9 ± 0.4 66.9 ± 0.4 66.2 ± 0.4 0.001 c,e
Obesity (%) < 0.001 a,b,c 0.002 a,c
Underweight 6.2 ± 1.4 1.6 ± 0.6 1.4 ± 0.5 1.0 ± 0.6 2.5 ± 0.7 0.9 ± 0.4 2.3 ± 0.7 1.8 ± 0.7
(< 18.5 kg/m2)
Normal (18.5–24.9 kg/m2) 73.3 ± 2.5 65.7 ± 3.3 59.3 ± 2.8 61.4 ± 3.1 63.3 ± 2.7 55.2 ± 2.8 55.1 ± 2.8 48.4 ± 2.8
Obese (≥25.0 kg/m2) 20.5 ± 2.3 32.7 ± 3.4 39.3 ± 2.9 37.6 ± 3.1 34.2 ± 2.8 43.8 ± 2.9 42.6 ± 2.8 49.8 ± 2.8
Ever-smoker (%) 52.1 ± 3.0 60.9 ± 2.9 55.5 ± 3.3 57.7 ± 3.5 0.618 – 5.8 ± 1.1 5.3 ± 1.1 8.6 ± 1.6 6.5 ± 1.4 0.203 –
Alcohol drinking (%) 40.3 ± 3.0 39.8 ± 5.5 47.1 ± 3.0 48.1 ± 3.7 0.271 – 10.1 ± 2.5 7.4 ± 1.5 8.4 ± 2.3 8.8 ± 2.1 0.616 –
Regular exercise (%) 41.4 ± 3.0 45.6 ± 3.2 49.8 ± 3.1 54.5 ± 3.0 0.003 b,c,e 36.6 ± 2.3 43.9 ± 2.6 48.8 ± 2.5 55.3 ± 2.4 0.001 b,c,e
Walking difficulty (%) 13.7 ± 1.4 11.5 ± 1.6 8.1 ± 1.2 7.6 ± 1.0 0.011 b,c 26.2 ± 2.2 22.6 ± 1.9 20.4 ± 1.9 15.8 ± 1.7 0.110 –
Household income (%) < 0.001 b,c,d 0.463
1st quartile 40.4 ± 2.7 35.0 ± 3.3 22.1 ± 2.4 29.0 ± 3.0 47.0 ± 3.1 36.8 ± 2.8 43.7 ± 2.9 41.2 ± 3.0

3
2nd quartile 27.1 ± 2.6 24.1 ± 2.8 27.1 ± 2.5 23.1 ± 2.5 23.7 ± 2.5 25.2 ± 2.3 24.2 ± 2.2 26.0 ± 2.4
3rd quartile 17.7 ± 2.1 24.5 ± 3.0 24.3 ± 2.3 24.6 ± 3.2 16.1 ± 2.3 20.9 ± 2.3 16.5 ± 1.9 19.7 ± 2.5
4th quartile 14.8 ± 1.9 16.4 ± 2.6 26.5 ± 2.7 23.2 ± 3.5 13.2 ± 1.9 17.1 ± 2.4 15.6 ± 2.2 13.1 ± 2.6
Total calorie intake (g/day) 2015 ± 48 2077 ± 55 2162 ± 45 2077 ± 45 0.124 – 1550 ± 28 1594 ± 49 1503 ± 30 1454 ± 32 0.081 –
Daily calcium intake (g/day) 497.9 ± 19.1 577.6 ± 24.8 564.7 ± 19.3 547.2 ± 19.7 0.036 a,b 390.2 ± 13.1 421.3 ± 20.5 401.2 ± 16.3 379.1 ± 13.9 0.351 –
Daily iron intake (g/day) 15.8 ± 0.7 17.1 ± 0.9 17.3 ± 0.7 16.6 ± 0.7 0.368 – 12.5 ± 0.4 14.0 ± 0.7 12.5 ± 0.5 13.8 ± 1.0 0.159 –
25(OH)D (ng/mL) 20.5 ± 0.5 20.7 ± 0.5 21.4 ± 0.5 20.7 ± 0.5 0.405 – 18.1 ± 0.4 18.5 ± 0.6 18.0 ± 0.5 18.6 ± 0.4 0.610 –
Alkaline phosphatase (IU/L) 251.8 ± 4.1 241.7 ± 5.4 236.6 ± 3.4 249.4 ± 4.7 0.020 b 256.2 ± 4.7 254.5 ± 4.2 263.4 ± 4.2 273.2 ± 4.5 0.017 c, e
PTH (pg/mL) 68.3 ± 2.0 66.2 ± 1.8 66.3 ± 1.5 65.8 ± 1.5 0.758 – 70.3 ± 1.7 70.6 ± 1.9 71.0 ± 1.8 71.9 ± 1.7 0.893 –
T-score
Femur neck −0.991 ± 0.053 −0.887 ± 0.066 −0.703 ± 0.055 −0.648 ± 0.062 < 0.001 b,c −1.706 ± 0.054 −1.634 ± 0.065 −1.533 ± 0.055 −1.461 ± 0.047 0.008 b,c,e
Spine −0.727 ± 0.079 −0.689 ± 0.087 −0.586 ± 0.063 −0.493 ± 0.0.075 0.023 c,e −1.877 ± 0.058 −1.868 ± 0.071 −1.688 ± 0.065 −1.539 ± 0.064 0.010 c,e

Values are expressed as mean ± SE percentage ± SE.

P-values were determined by weighted ANOVA or x2 test.
Post hoc analysis of weighted ANOVA or chi-squared test for mean between-group difference.
a. Q1 versus Q2; b. Q1 versus Q3; c. Q1 versus Q4; d. Q2 versus Q3; e. Q2 versus Q4; f. Q3 versus Q4.
Abbreviations: BMI, body mass index; ALP, alkaline phosphatase; PTH, parathyroid hormone; BMD, bone mineral density.
Experimental Gerontology 126 (2019) 110706
H.-S. Kim, et al.
statistical analysis, we applied sampling weights to account for complex
sampling. Because of the survey's characteristics, we presented the re-
sults as standard error than standard deviation. Clinical and chemical
characteristics of the study population according to the hemoglobin
quartiles were compared using weighted one-way analysis of variance
(ANOVA) for continuous variables and weighted chi-square test for
categorical variables. The mean differences and proportion differences
between groups were determined using post hoc analysis of weighted
ANOVA test and chi-squared test using Bonferroni corrections. In
multiple logistic regression analysis, the ORs for low BMD in both men
and women were calculated after adjusting for confounding variables
across hemoglobin quartiles. All analyses were conducted using SAS
statistical software, version 9.4 (SAS Institute Inc., Cary, NC, USA). All
statistical tests were two-tailed and a P-value of < 0.05 was considered
statistically significant.
3. Results
Table 1 shows the characteristics of the participants in relation to
hemoglobin quartiles. For both men and women, the mean values of age
and T-score were lowest and the proportion of underweight and
walking difficulty was lowest in the 4th quartile of both men and
women. Furthermore, the proportion of those who were underweight
and had difficulty in walking was lowest, whereas the proportion of
those who took regular exercise was highest in the 4th quartile of he-
moglobin. For men, daily dietary calcium intake was lowest, whereas
serum ALP level was highest, and the proportion of the lowest level of
household income in the 1st quartile of hemoglobin. For women, serum
ALP level was highest in the 4th quartile of hemoglobin.
Fig. 1 presents the prevalence of low BMD according to hemoglobin
quartiles. The prevalence of low BMD in the femur and spine gradually
decreased in accordance with the hemoglobin quartiles of men and
women.
Fig. 2 shows the results of multiple logistic regression analysis to
assess the odds for predicting low BMD in terms of hemoglobin quar-
tiles. Compared with the group in the lowest quartile, the OR (95% CI)
for low BMD in the lumbar spine was 0.78 (0.54–0.93) for men and 0.67
(0.50–0.93) for women after adjusting for age, BMI, tobacco smoking,
alcohol intake, physical activity, walking difficulty, household income,
total calorie intake, calcium intake, iron intake, 25(OH)D, alkaline
phosphatase, and parathyroid hormone levels. However, these inverse
associations were not observed for the femoral neck in both sexes.
Fig. 1. Prevalence of low BMD according to hemoglobin quartiles in men and women. Post hoc analysis of weighted chi-squared test for mean between-group
difference (a. Q1 versus Q2; b. Q1 versus Q3; c. Q1 versus Q4; d. Q2 versus Q3; e. Q2 versus Q4; f. Q3 versus Q4).
4
Experimental Gerontology 126 (2019) 110706
4. Discussion
In this nationally representative cross-sectional study, we in-
vestigated the relationship between hemoglobin levels and low BMD in
a non-anemic healthy elderly population. We found that hemoglobin
levels were inversely associated with low BMD in the lumbar spine
among non-anemic healthy elderly.
Our results are in agreement with previous studies on the positive
associations between hemoglobin levels and BMD (Jorgensen et al.,
2010). Laudisio et al. reported that hemoglobin levels were positively
associated with BMD as measured by ultrasound in 358 elderly Tuscan
individuals (173 females with normal hemoglobin and 24 anemic fe-
males), and also that anemia was a risk factor for low BMD (Laudisio
et al., 2009). In an Italian cohort study, anemia was negatively asso-
ciated with bone density assessed by tibia peripheral quantitative
computed tomography (pQCT) (Cesari et al., 2005; Korkmaz et al.,
2012). However, the BMD measurements in these previous studies were
based on a non-axial peripheral BMD measurement method or de-
termined by ultrasound. Whereas ISCD and the WHO recommend
central DXA in axial bones such as the lumbar spine, total hip, or fe-
moral neck. In another study of 371 postmenopausal Turkish women,
Korkmaz et al. reported that anemia is an independent predictor of low
bone mass after adjusting for BMI, smoking, age at menarche, physical
activity, and duration of menopause (Korkmaz et al., 2012). However,
the lack of measurement of biochemical indices of bone mineral me-
tabolism including alkaline phosphatase, parathyroid hormone, and
25(OH)D level was a crucial limitation of that study. In addition, it
remains controversial as to whether osteoporosis is an epiphenomenon
of anemia accompanied by malabsorption and miscellaneous entities
(Onal and Usluogullari, 2012). To overcome the limitations of previous
studies, we excluded participants who had anemia and investigated
only healthy individuals who met our inclusion criteria. Moreover, we
included data on the confounding biochemical factors alkaline phos-
phatase, parathyroid hormone, and 25(OH)D levels in the multiple lo-
gistic regression analyses models.
Although a biological pathway linking the inverse relationship be-
tween hemoglobin levels and low BMD has not yet been fully estab-
lished, there are several plausible explanations. First, the key concept of
a possible mechanism between hemoglobin and BMD may be the age-
associated conversion of red to yellow bone marrow in the axial ske-
leton. Specifically, bone marrow consists of red marrow, which is made
up of hematopoietic tissue, and yellow marrow, which is mainly
H.-S. Kim, et al. Experimental Gerontology 126 (2019) 110706

Fig. 2. Odds ratios 95% confidence intervals for low BMD according to hemoglobin quartiles in men (A) and women (B). Multiple logistic regression analyses were
conducted after adjusting for age, body mass index, tobacco smoking, alcohol intake, physical activity, walking difficulty, household income, total calorie intake,
calcium intake, and iron intake, serum levels of 25-hydroxyvitamin D, alkaline phosphatase, and parathyroid hormone.

composed of fat cells. Red blood cells, platelets, and most subtypes of
leukocytes are produced in the red marrow (Gurevitch et al., 2007). In
this way, the age-related conversion of red to yellow bone marrow in
the axial skeleton may explain how hemoglobin levels are inversely
associated with low BMD (Schellinger et al., 2004; Schwartz et al.,
2013). Moreover, spine bone is composed of a 3:1 ratio of trabecular
and cortical bone, whereas this ratio is 1:1 in the femoral bone. The
trabecular bone turnover rate is known to be about eight- to ten-fold
higher than that of cortical bone, because trabecular bone contains
most of the body's red bone marrow, i.e. hematopoietic stem cells
(Clarke, 2008). In this regard, lumbar spine was thought to more sen-
sitively reflect the loss of red marrow than femoral neck in the present
study.
Secondly, relative hypoxia accompanied by decreased hemoglobin
levels may have an adverse effect on bone metabolism via the direct
action of reduced pO2 and an indirect effect on bone formation and
resorption cells (Bozec et al., 2008; Knowles and Athanasou, 2009;
Utting et al., 2010). Oxygen tension has long been considered to be
important in skeleton homeostasis. Published data regarding the effects
of hypoxia on bone marrow mesenchymal stem cells include
stimulation of fibroblast proliferation, increased osteogenic potential,
and stimulation of adipogenesis (Malladi et al., 2007). A recent rodent
study showed that hypoxia blocks the growth and differentiation of
osteoblasts while strongly stimulating osteoclast formation (Arnett,
2010). Because the majority of the body's oxygen supply is transported
by hemoglobin, the inverse association between hemoglobin levels and
low BMD in healthy elderly individuals may be explained by the hy-
poxia mechanism.
There are some limitations of the present that should be considered
when interpreting its findings. First, this was a cross-sectional study by
design and thus it was not possible to establish a causal relationship
between hemoglobin levels and BMD. Further prospective studies with
a larger number of participants over a longer period of time are war-
ranted to confirm the cause and effect relationship of hemoglobin levels
and BMD. Second, hemoglobin levels in this study were only measured
once, and thus it was not possible to determine whether acute or brief
episodes of low hemoglobin levels were responsible for the observed
correlations. Despite these potential limitations, the inverse association
between hemoglobin levels and BMD reported in this paper are based
on nationally representative data and may be useful for the clinical

5
H.-S. Kim, et al.
setting. Moreover, we believe that the present study is the first to
specifically examine the relationship between hemoglobin levels and
BMD in a non-anemic healthy elderly population.
5. Conclusions
In conclusion, hemoglobin levels were inversely associated with low
BMD the lumbar spine among non-anemic elderly individuals.
Author contributions
H-S Kim designed the study, wrote and edited the manuscript. Y-J
Lee and H-S Lee did statistical analyses and interpreted the data. H-M
Park, and Y-J Lee contributed to the discussion and revised the manu-
script. Y-J Lee, as the corresponding author, coordinated the study,
interpreted the data, contributed to the discussion, and wrote the
manuscript. All authors read and approved the final version.
Declaration of competing interest
No potential conflict of interest was reported by the authors.
Acknowledgements
We thank all those who conducted the KNHANES as well as the
civilians who participated in this survey.
Funding information
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
References
Arnett, T.R., 2010. Acidosis, hypoxia and bone. Arch. Biochem. Biophys. 503, 103–109.
https://doi.org/10.1016/j.abb.2010.07.021.
Bozec, A., Bakiri, L., Hoebertz, A., Eferl, R., Schilling, A.F., Komnenovic, V., Scheuch, H.,
Priemel, M., Stewart, C.L., Amling, M., Wagner, E.F., 2008. Osteoclast size is con-
trolled by Fra-2 through LIF/LIF-receptor signalling and hypoxia. Nature 454,
221–225. https://doi.org/10.1038/nature07019.
Cesari, M., Pahor, M., Lauretani, F., Penninx, B.W., Bartali, B., Russo, R., Cherubini, A.,
Woodman, R., Bandinelli, S., Guralnik, J.M., Ferrucci, L., 2005. Bone density and
hemoglobin levels in older persons: results from the InCHIANTI study. Osteoporosis
international: a journal established as result of cooperation between the European
Foundation for Osteoporosis and the National Osteoporosis Foundation of the USA
16, 691–699. https://doi.org/10.1007/s00198-004-1739-6.
Clarke, B., 2008. Normal bone anatomy and physiology. Clinical journal of the American
Society of Nephrology: CJASN 3 (Suppl. 3), S131–S139. https://doi.org/10.2215/cjn.
04151206.
Duh, M.S., Mody, S.H., Lefebvre, P., Woodman, R.C., Buteau, S., Piech, C.T., 2008.
Anaemia and the risk of injurious falls in a community-dwelling elderly population.
Drugs Aging 25, 325–334. https://doi.org/10.2165/00002512-200825040-00005.
Gurevitch, O., Slavin, S., Feldman, A.G., 2007. Conversion of red bone marrow into
yellow - cause and mechanisms. Med. Hypotheses 69, 531–536. https://doi.org/10.
1016/j.mehy.2007.01.052.
Ha, Y.-C., 2016. Epidemiology of osteoporosis in Korea. J Korean Med Assoc 59, 836–841.
Jorgensen, L., Skjelbakken, T., Lochen, M.L., Ahmed, L., Bjornerem, A., Joakimsen, R.,
Jacobsen, B.K., 2010. Anemia and the risk of non-vertebral fractures: the Tromso
Experimental Gerontology 126 (2019) 110706
study. Osteoporosis international: a journal established as result of cooperation be-
tween the European Foundation for Osteoporosis and the National Osteoporosis
Foundation of the USA 21, 1761–1768. https://doi.org/10.1007/s00198-009-
1131-7.
Kelly, T.L., Wilson, K.E., Heymsfield, S.B., 2009. Dual energy X-ray absorptiometry body
composition reference values from NHANES. PLoS One 4, e7038. https://doi.org/10.
1371/journal.pone.0007038.
Knowles, H.J., Athanasou, N.A., 2009. Acute hypoxia and osteoclast activity: a balance
between enhanced resorption and increased apoptosis. J. Pathol. 218, 256–264.
https://doi.org/10.1002/path.2534.
Korkmaz, U., Korkmaz, N., Yazici, S., Erkan, M., Baki, A.E., Yazici, M., Ozhan, H.,
Ataoglu, S., 2012. Anemia as a risk factor for low bone mineral density in post-
menopausal Turkish women. European journal of internal medicine 23, 154–158.
https://doi.org/10.1016/j.ejim.2011.11.009.
Laudisio, A., Marzetti, E., Pagano, F., Bernabei, R., Zuccala, G., 2009. Haemoglobin levels
are associated with bone mineral density in the elderly: a population-based study.
Clin. Rheumatol. 28, 145–151. https://doi.org/10.1007/s10067-008-0998-6.
Leng, S., Chaves, P., Koenig, K., Walston, J., 2002. Serum interleukin-6 and hemoglobin
as physiological correlates in the geriatric syndrome of frailty: a pilot study. J. Am.
Geriatr. Soc. 50, 1268–1271.
Malladi, P., Xu, Y., Chiou, M., Giaccia, A.J., Longaker, M.T., 2007. Hypoxia inducible
factor-1alpha deficiency affects chondrogenesis of adipose-derived adult stromal
cells. Tissue Eng. 13, 1159–1171. https://doi.org/10.1089/ten.2006.0265.
Onal, E.D., Usluogullari, A., 2012. Anemia and osteoporosis: causal association or epi-
phenomenon? European journal of internal medicine 23, e117. https://doi.org/10.
1016/j.ejim.2012.03.010.
Organization, W.H., 1968. Nutritional Anaemias: Report of a WHO Scientific Group
[Meeting Held in Geneva From 13 to 17 March 1967].
Papadimitriou, N., Tsilidis, K.K., Orfanos, P., Benetou, V., Ntzani, E.E., Soerjomataram, I.,
Kunn-Nelen, A., Pettersson-Kymmer, U., Eriksson, S., Brenner, H., Schottker, B.,
Saum, K.U., Holleczek, B., Grodstein, F.D., Feskanich, D., Orsini, N., Wolk, A.,
Bellavia, A., Wilsgaard, T., Jorgensen, L., Boffetta, P., Trichopoulos, D.,
Trichopoulou, A., 2017. Burden of hip fracture using disability-adjusted life-years: a
pooled analysis of prospective cohorts in the CHANCES consortium. Lancet Public
Health 2, e239–e246. https://doi.org/10.1016/s2468-2667(17)30046-4.
Reginster, J.Y., Burlet, N., 2006. Osteoporosis: a still increasing prevalence. Bone 38,
S4–S9. https://doi.org/10.1016/j.bone.2005.11.024.
Rutten, E.P., Franssen, F.M., Spruit, M.A., Wouters, E.F., 2013. Anemia is associated with
bone mineral density in chronic obstructive pulmonary disease. Copd 10, 286–292.
https://doi.org/10.3109/15412555.2012.744390.
Salive, M.E., Cornoni-Huntley, J., Guralnik, J.M., Phillips, C.L., Wallace, R.B., Ostfeld,
A.M., Cohen, H.J., 1992. Anemia and hemoglobin levels in older persons: relationship
with age, gender, and health status. J. Am. Geriatr. Soc. 40, 489–496.
Sarrai, M., Duroseau, H., D’Augustine, J., Moktan, S., Bellevue, R., 2007. Bone mass
density in adults with sickle cell disease. Br. J. Haematol. 136, 666–672. https://doi.
org/10.1111/j.1365-2141.2006.06487.x.
Schellinger, D., Lin, C.S., Lim, J., Hatipoglu, H.G., Pezzullo, J.C., Singer, A.J., 2004. Bone
marrow fat and bone mineral density on proton MR spectroscopy and dual-energy X-
ray absorptiometry: their ratio as a new indicator of bone weakening. AJR Am. J.
Roentgenol. 183, 1761–1765. https://doi.org/10.2214/ajr.183.6.01831761.
Schoeller, D.A., Tylavsky, F.A., Baer, D.J., Chumlea, W.C., Earthman, C.P., Fuerst, T.,
Harris, T.B., Heymsfield, S.B., Horlick, M., Lohman, T.G., Lukaski, H.C., Shepherd, J.,
Siervogel, R.M., Borrud, L.G., 2005. QDR 4500A dual-energy X-ray absorptiometer
underestimates fat mass in comparison with criterion methods in adults. Am. J. Clin.
Nutr. 81, 1018–1025. https://doi.org/10.1093/ajcn/81.5.1018.
Schwartz, A.V., Sigurdsson, S., Hue, T.F., Lang, T.F., Harris, T.B., Rosen, C.J., Vittinghoff,
E., Siggeirsdottir, K., Sigurdsson, G., Oskarsdottir, D., Shet, K., Palermo, L.,
Gudnason, V., Li, X., 2013. Vertebral bone marrow fat associated with lower trabe-
cular BMD and prevalent vertebral fracture in older adults. J. Clin. Endocrinol.
Metab. 98, 2294–2300. https://doi.org/10.1210/jc.2012-3949.
Taal, M.W., Masud, T., Green, D., Cassidy, M.J., 1999. Risk factors for reduced bone
density in haemodialysis patients. Nephrology, dialysis, transplantation: official
publication of the European Dialysis and Transplant Association - European Renal
Association 14, 1922–1928. https://doi.org/10.1093/ndt/14.8.1922.
Utting, J.C., Flanagan, A.M., Brandao-Burch, A., Orriss, I.R., Arnett, T.R., 2010. Hypoxia
stimulates osteoclast formation from human peripheral blood. Cell Biochem. Funct.
28, 374–380. https://doi.org/10.1002/cbf.1660.